ANALYTIC REVIEWS
Critical Care Management of Increased
Intracranial Pressure
Stephan A. Mayer, MD*²
and Ji Y. Chong, MD*
Mayer SA, Chong JY. Critical care management of increased
intracranial pressure. J Intensive Care Med 2002;17:55±67.
Increased intracranial pressure (ICP) is a pathologic state
common to a variety of serious neurologic conditions, all of
which are characterized by the addition of volume to the
intracranial vault. Hence all ICP therapies are directed toward
reducing intracranial volume. Elevated ICP can lead to brain
damage or death by two principle mechanisms: (1) global
hypoxic-ischemic injury, which results from reduction of
cerebral perfusion pressure (CPP) and cerebral blood ¯ow,
and (2) mechanical compression, displacement, and herni-
ation of brain tissue, which results from mass effect
associated with compartmentalized ICP gradients. In
unmonitored patients with acute neurologic deterioration,
head elevation (30 degrees), hyperventilation (pCO2 26±30
mmHg), and mannitol (1.0±1.5 g/kg) can lower ICP within
minutes. Fluid-coupled ventricular catheters and intraparen-
chymal pressure transducers are the most accurate and
reliable devices for measuring ICP in the intensive care unit
(ICU) setting. In a monitored patient, treatment of critical ICP
elevation (>20 mmHg) should proceed in the following steps:
(1) consideration of repeat computed tomography (CT)
scanning or consideration of de®nitive neurosurgical inter-
vention, (2) intravenous sedation to attain a quiet, motionless
state, (3) optimization of CPP to levels between 70 and 110
mmHg, (4) osmotherapy with mannitol or hypertonic saline,
(5) hyperventilation (pCO2 26±30 mmHg), (6) high-dose
pentobarbital therapy, and (7) systemic cooling to attain
moderate hypothermia (32±33°C). Placement of an ICP
monitor and use of a stepwise treatment algorithm are both
essential for managing ICP effectively in the ICU setting.
From the Division of Critical Care Neurology, Departments of
*Neurology and ²Neurosurgery, College of Physicians and
Surgeons, Columbia University, New York, NY.
Received Jul 30, 2001, and in revised form Aug 24, 2001. Accepted
for publication Aug 28, 2001.
Address correspondence to Dr. Mayer, Neurological Institute,
710 West 168th St., Unit 39, New York, NY 10032, or e-mail:
sam14@columbia.edu.
Increased intracranial pressure (ICP) can result
from a number of insults to the brain, including
traumatic brain injury (TBI), stroke, encephalitis,
neoplasms, and abscesses (Table 1). The funda-
mental abnormality common to these diverse
disease states is an increase in intracranial volume.
Accordingly, all treatments for elevated ICP work by
reducing intracranial volume. Prompt recognition
and treatment of elevated ICP is essential because
sustained elevated ICP can cause brain damage or
be rapidly fatal.
Pathophysiology
Intracranial Compliance. The Monroe±Kellie
doctrine dictates that the cranial vault is a ®xed
space that contains three compartments: blood,
cerebrospinal ¯uid (CSF), and brain tissue. In the
average adult, the brain volume is 1400 ml, the
blood volume is 150 ml, and the CSF volume is 150
ml. CSF is produced by the choroid plexus in the
ventricles at a rate of approximately 20 ml/hr, and
drains into the venous system via the arachnoid villi
and granulations [1]. This out¯ow is normally of low
resistance; hence jugular venous pressure is the
chief determinant of ICP in healthy patients. Normal
ICP ranges from 50 to 200 mmH2O or 3±15 mmHg.
In routine intensive care unit (ICU) practice, the
goal of ICP management is to maintain levels below
20 mmHg.
In pathologic states characterized by increased
ICP (Table 1), additional volume is added to the
intracranial compartment. This can result from the
addition of an extrinsic mass lesion or from an
increase in the volume of CSF (hydrocephalus),
brain tissue (cytotoxic edema), or blood (vasogenic
edema). To maintain ICP within normal limits, these
increases in intracranial volume are initially coun-
terbalanced by volume reductions in the other
compartments. CSF is displaced through the fora-
men magnum into the paraspinal space, blood is
displaced from the intracranial to the extracranial
venous system, and the brain parenchyma is com-
pressed. After these mechanisms are exhausted,
Copyright Ó 2002 Blackwell Science, Inc. 55
56 Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

Table 1. Conditions Associated with Increased ICP

Intracranial mass lesions

Subdural hematoma
Epidural hematoma
Brain tumor
Cerebral abscess
Intracerebral hemorrhage
Increased brain volume (cytotoxic edema)
Cerebral infarction
Global hypoxia-ischemia
Reye's syndrome
Acute hyponatremia
Increased blood and brain volume (vasogenic edema)
Hepatic encephalopathy
Traumatic brain injury Fig 1. Intracranial pressure-volume relationship. At
Meningitis point A, on the ¯at portion of the curve, the amplitude of
Encephalitis the arterial re¯ection in the ICP waveform is small (inset),
Hypertensive encephalopathy and the addition of volume leads to a small increase, in
Eclampsia pressure (A¢). At point B, on a steeper portion of the curve,
Subarachnoid hemorrhage the intracranial compartment is relatively noncompliant,
Dural sinus thrombosis the amplitude of the arterial re¯ection in the ICP wave-
Altitude-related cerebral edema form is large (inset), and addition of the same amount of
Increased CSF volume volume leads to a larger increase, in pressure (B¢). (Re-
Communicating hydrocephalus printed from Mayer SA. Management of increased intra-
Noncommunicating hydrocephalus cranial pressure. In: Wijdicks EFM, Diringer MN, Bolton
Choroid plexus papilloma CF, et al. Continuum: Critical Care. Minneapolis, MN:
American Academy of Neurology, 1997:47±61.)
intracranial compliance (DV/DP) falls sharply, and
even small increases in intracranial volume can lead
to dramatic elevations in ICP (Fig 1).
The relative state of intracranial compliance can
be assessed by inspection of the ICP waveform
(Fig 1, insets). ICP normally increases 2±3 mmHg
with each arterial pulse because of transient increa-
ses in cerebral blood volume (CBV). When intra-
cranial compliance is reduced, intracranial pulse
pressure can reach levels of 10±15 mmHg, which
re¯ects loss of the ability to accommodate even
small pulsatile increases in CBV. As intracranial
compliance falls, the morphology of the ICP
waveform also changes, in that the amplitude of
the second peak (the dicrotic wave) initially equals
and then exceeds the amplitude of the ®rst peak
(the percussion wave) (Fig 2).
Fig 2. ICP waveform in conditions of normal (top) and
abnormal (bottom) intracranial compliance. (Reprinted
Cerebral Perfusion Pressure. Cerebral perfu- from Chestnut RM, Marshall LF. Treatment of abnormal
sion pressure (CPP), de®ned as the mean arterial intracranial pressure. Neurosurg Clin N Am 1991;2:267±
pressure (MAP) minus ICP, is a critical determinant 284.)
of cerebral blood ¯ow (CBF) and plays an import-
ant role in ICP management. Normally CBF is subarachnoid hemorrhage, CBF may approximate a
``autoregulated'' at a constant level over a wide linear relationship with CPP, which creates a
range of CPPs (from 50 to 150 mmHg) (Fig 3). smaller range of optimal CPP (Fig 2). Reduction of
Pressure autoregulation of this type is mediated by CPP below the lower limit of autoregulation can
changes in arteriolar diameter and cerebrovascular lead to ischemia [2], whereas CPP elevation above
resistance. The autoregulatory curve is shifted to the the upper limit of autoregulation can be associated
left in children and shifted to the right in patients with hyperemia, exacerbation of vasogenic edema,
with chronic hypertension. In pathologic states and increased ICP [3]. Although the optimal CPP for
with impaired autoregulation, such as TBI and any particular patient may vary, as a rule of thumb

Fig 3. Cerebral autoregulation curve. In the normal
relationship (solid line), with CBF held constant across a
wide range of CPP (50±150 mmHg). In disease states (e.g.,
vasospasm, ischemia, intracranial mass lesion), cerebral
blood ¯ow may become pressure passive (dotted line).
(Reprinted from Mayer SA. Management of increased
intracranial pressure. In: Wijdicks EFM, Diringer MN,
Bolton CF, et al. Continuum: Critical Care. Minneapolis,
MN: American Academy of Neurology, 1997:47±61.)
CPP should be maintained above 70 mmHg to avert
ischemia, and below 110 mmHg to avoid break-
through hyperperfusion. To accurately measure
CPP in the ICU, the pressure transducer used to
measure MAP must be set at head level [4].
CBF also depends upon PaCO2 and PaO2 levels.
In general, the cerebral vessels are less responsive
to changes in PaO2 than to those in PaCO2.
Arteriolar diameter and CBF progressively increase
as PaCO2 rises from 20 to 80 mmHg, whereas
hypoxemia leads to vasodilation and increased CBF
only when PaO2 falls below 50 mmHg [1].
Pathologic ICP Waves. Patients with reduced
intracranial compliance and elevated ICP may
develop pathologic ICP waves (Fig 4). Lundberg
A waves (plateau waves) are dangerous elevations
in ICP [5]. They occur suddenly, can reach levels of
50±100 mmHg, and can last from minutes to hours.
Plateau waves are characteristically associated with
``mirror'' reductions in CPP (Fig 5). When severe,
plateau waves are associated with reduced CPP and
CBF, leading to global hypoxic-ischemic damage.
Lundberg B waves are of lesser amplitude (5±20
mmHg) and of shorter duration (1±5 minutes) than
A waves. Although they are not directly harmful, B
waves are a useful marker of abnormal intracranial
compliance. Both of these waves characteristically
end with a surge in systemic blood pressure and
ICP, known as a termination spike [5,6].
A vasodilatory cascade model has been proposed
to explain the pathogenesis of pathologic ICP
waves [6,7]. According to this model, pathologic A
and B waves occur because CPP is inadequate. The
Mayer and Chong: Management of ICP 57
Fig 4. Pathologic ICP elevations. (A) Lundberg A (plat-
eau) waves. (B) Lundberg B waves. (Reprinted from
Mayer SA. Management of increased intracranial pres-
sure. In: Wijdicks EFM, Diringer MN, Bolton CF, et al.
Continuum: Critical Care. Minneapolis, MN: American
Academy of Neurology, 1997:47±61.)
process begins with a reduction in CPP, which
results from a drop in MAP or a surge in ICP [6]. To
maintain CBF, the cerebral vasculature dilates and
CBV increases. This adds volume to the intracranial
compartment, resulting in an increase in ICP and
further reduction of CPP. This initiates a vicious
cycle in which cerebral vasodilation continues until
it is maximal, at which point a ``plateau'' is reached
at a new level of increased CBV and ICP and
decreased CPP and CBF. The plateau ends once
CBF is inadequate to maintain tissue oxygenation
and ischemia develops. This results in a re¯ex
systemic pressor response mediated by a surge in
systemic vascular resistance. MAP climbs and CPP is
restored (the termination spike), which allows the
Fig 5. Plateau waves followed by sustained ICP eleva-
tion in a patient with traumatic brain injury. Early in the
course of the recording, plateau waves exceeding 90
mmHg are associated with ``mirror'' reductions in CPP
below 50 mmHg. At the end of the recording, ICP remains
elevated between 40 and 60 mmHg, MAP falls below 75
mmHg, and CPP drops to 20 mmHg. At this point the
patient became clinically brain dead.
58 Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

cerebral vessels to return to normal caliber, restor-
ing CBV and ICP to normal levels.
Signs of Increased ICP and Herniation
The clinical manifestations of increased ICP are well
known, but are notoriously unreliable (Table 2). A
depressed level of consciousness and re¯ex hyper-
tension, the two most consistent signs, both re¯ect
the effects of globally reduced CBF. However, many
patients have multiple reasons for a depressed level
of consciousness, and in some patients with signi®-
cant shift and mass effect, ICP may be normal.
Comatose patients with intracerebral hemorrhage
[8] and middle cerebral artery territory infarction [9],
for instance, may have ICP levels that vary from
normal to highly elevated, and brain stem hernia-
tion can occur in the absence of elevated ICP.
Cushing's triad (hypertension, bradycardia, and
irregular respiration), which was originally des-
cribed in response to elevated ICP, can also result
from brain stem herniation. Because of the poor
correlation between clinical signs and ICP, the only
way to properly diagnose increased ICP is to
directly measure it.
It is important to differentiate clinical signs of
increased ICP from signs of cerebral herniation
(Table 3). Brain tissue displacement and herniation
occur when compartmentalized mass effect leads to
ICP gradients [10,11]. Patients with intracranial mass
Table 2. Clinical Signs of Increased ICP
Signs which are almost always present
Depressed level of consciousness
(lethargy, stupor, oma)
Hypertension, with or without bradycardia
Symptoms and signs which are sometimes present
Headache
Vomiting
Papilledema
Sixth cranial nerve palsies
lesions may have elevated ICP, brain tissue shifting,
or both. Herniation is often rapidly fatal, but can be
reversed by reducing mass effect related to
compartmentalized ICP gradients with treatments
such as mannitol, hypertonic saline, and hyperven-
tilation.
ICP Monitoring
Indications. Invasive monitoring of ICP is gener-
ally indicated in patients who meet all three of the
following criteria:
1. The patient is suspected to be at risk for elevated
ICP.
2. The patient is comatose (Glasgow coma scale
score £ 8).
3. The prognosis is such that aggressive ICU treat-
ment is indicated.
Suspicion of increased ICP is usually based on
clinical signs (Tables 2 and 3) and the results of a
computed tomography (CT) scan showing signi®-
cant intracranial mass effect with midline shift or
effacement of the basal cisterns. However, in
comatose patients with TBI, intracranial hyperten-
sion occurs in approximately 10% of patients with
normal CT scans; this risk is even higher in patients
more than 40 years old, with motor posturing, or
with hypotension (systolic blood pressure < 90
mmHg) [12].
If a patient is awake and can follow commands, it
is unlikely that ICP is dangerously elevated [13], and
the bene®ts of ventricular drainage or ICP monit-
oring probably do not outweigh the risks. Careful
monitoring of mental status in an ICU will usually
suf®ce in these cases.
Invasive ICP monitoring devices. Empiric ther-
apy for increased ICP (i.e., standing doses of mann-
itol) without invasive monitoring is a distressingly
common practice. This approach is unsatisfactory

Table 3. Herniation Syndromes

Type Clinical hallmark Causes

Uncal (lateral transtentorial) Ipsilateral CN 3 palsy Temporal lobe mass lesion

Contralateral motor posturing
Central transtentorial Coma with progression from bilateral Diffuse cerebral edema
decorticate to decerbrate posturing
Rostral-caudal loss of brain stem re¯exes Acute hydrocephalus
Subfalcine Coma with asymmetric Convexity (frontal or parietal)
(contralateral > ipsilateral) motor posturing mass lesion
Cerebellar Sudden progression to coma with bilateral Cerebellar mass lesion
(upward or downward) motor posturing in a patient with cerebellar signs

because most ICP treatments are effective for a short
time only, may lose their ef®cacy with prolonged
use, and have side effects. Optimally therapy
should be given when ICP is high, and withheld
when it is normal. Only an invasive ICP monitor
makes this possible. The four main types of invasive
ICP monitors used in standard ICU practice are
listed below, in general order of their accuracy and
reliability (Fig 6).
INTRAVENTRICULAR CATHETERS. These devices directly
connect the intracranial space to an external
pressure transducer via saline-®lled tubing. The
bedside pressure transducer must be positioned at
the level of the foramen of Monroe (external
auditory meatus) to accurately re¯ect ICP. The
catheter is usually connected to both a pressure
transducer and an external drainage system via a
three-way stopcock. The system can then be set for
continuous ICP monitoring with intermittent CSF
drainage or continuous drainage with intermittent
ICP measurement. The major advantage to intra-
ventricular catheters (IVCs) is that they allow
treatment of increased ICP via drainage of CSF.
The main disadvantage is the high risk of infection
(ventriculitis or meningitis), which occurs in
10±20% of patients and increases dramatically after
5 days [14]. Most neurointensivists administer
prophylactic antibiotics with gram-positive cover-
age, such as oxacillin 1±2 g every 6 hours, to
minimize this risk. In a clinical trial of 228 IVC
patients, prolonged therapy with ampicillin sulbac-
tam in the ICU signi®cantly reduced the frequency
of CSF infection compared to patients given peri-
operative treatment only (3% versus 11%) [15].
Fig 6. ICP monitoring devices. (A) Intraparenchymal
®beroptic probe or microsensor; (B) intraventricular
catheter; (C) epidural transducer; (D) subarachnoid bolt.
(Reprinted from Mayer SA. Management of increased
intracranial pressure. In: Wijdicks EFM, Diringer MN,
Bolton CF, et al. Continuum: Critical Care. Minneapolis,
MN: American Academy of Neurology, 1997:55.)
Mayer and Chong: Management of ICP 59
INTRAPARENCHYMAL PRESSURE TRANSDUCERS. The pres-
sure transducer in these disposable devices is
incorporated into the tip of a thin ®beroptic cable
(the Camino device) or within a strain-gauge
microsensor at the tip of a ¯exible catheter (the
Codman device). These catheters can be placed into
either the brain parenchyma or the ventricle via a
small burr hole and screw [16,17]. With intraparen-
chymal placement, the infection rate is exceedingly
low (approximately 1%) [18]. When combined with
a ventricular catheter, the system allows simulta-
neous CSF drainage and continuous ICP measure-
ment. These devices only need to be calibrated
once prior to insertion, and the accuracy of ICP
measurements is generally superior to those provi-
ded by subarachnoid bolts or epidural transducers
[19]. A new version of the Codman monitor also
provides measurements of brain temperature. A
third parenchymal monitor recently approved by
U.S. Food and Drug Administration (FDA) (the
Spielberg device) features a small air-®lled balloon
at the tip of a ¯exible catheter; it has the advantage
of providing measurements of intracranial compli-
ance (calculated as a pressure/volume index) as
well as ICP [20].
SUBARACHNOID BOLTS. This is another ¯uid-coupled
system which connects the intracranial space to an
external transducer at the bedside via saline-®lled
tubing [21]. The subarachnoid bolt is actually a
hollow screw that is inserted via a burr hole. The
dura at the base of the bolt is perforated with a
spinal needle, allowing the subarachnoid CSF to ®ll
the bolt. Pressure tubing is then connected to
establish communication with a pressure monitor-
ing system. Although the infection risk is low, these
devices are prone to error, including underestima-
tion of ICP, screw displacement, and occlusion by
debris [22].
EPIDURAL TRANSDUCERS. These devices (the Gaeltec
device) are inserted deep into the inner table of the
skull and super®cial to the dura [23]. In most of these
devices, pressure is transduced by an optical sensor.
They have a low infection rate (approximately 1%)
[17], but are prone to malfunction, displacement,
and baseline drift that can exceed 5±10 mmHg after
more than a few days of use. Much of the inaccuracy
results from having the relatively inelastic dura
between the sensor tip and the subarachnoid space.
Noninvasive ICP Monitoring. At present there is
no noninvasive method that can provide accurate
continuous on-line measurement of ICP. However,
transcranial Doppler (TCD) ultrasonography,
which measures the velocity of blood ¯ow in the
60 Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002
basal cerebral arteries, shows characteristic changes
with increasing ICP [24]. As CPP falls, diastolic
velocity decreases and pulsatility increases, re¯ect-
ing increased distal vascular resistance to ¯ow.
Though this ®nding is speci®c for severe intracra-
nial hypertension, TCD is not sensitive to mild to
moderate ICP elevations. Lateralized asymmetries
in TCD pulsatility correlate with lesion volume in
intracerebral hemorrhage, and are believed to
re¯ect compartmentalized ICP gradients [25]. Prom-
ising new applications using ultrasound technology
to estimate ICP noninvasively have been described,
but have not yet been validated in the clinical setting
[26±28].
Adjuncts to ICP Monitoring. Several new
modalities have recently been introduced that can
provide additional information regarding the ad-
equacy of cerebral perfusion and the extent of
injury in patients undergoing ICP monitoring.
JUGULAR VENOUS OXYGEN SATURATION (SJVO2) AND
BRAIN TISSUE PO2 (PBTO2) MONITORING. SjvO2 monit-
oring assesses the adequacy of global cerebral
oxygen delivery, whereas PbtO2 monitoring meas-
ures regional oxygen tension. Both emerging
technologies provide continuous information re-
garding the adequacy of CPP and CBF at the tissue
level [29±31]. SjvO2 is measured with a 5-French
®beroptic oxygen saturation catheter placed retro-
grade in the internal jugular vein so that the tip is
positioned in the jugular bulb; PbtO2 is measured
with a miniaturized Clark electrode embedded in
the tip of a thin catheter inserted 3±4 cm into the
cerebral white matter (the Licox or Neurotrend
device). Both techniques can detect inadequate
CBF (i.e., ischemia), which may occur even in
patients with relatively normal CPP [32], and
excessive CBF (i.e., hyperemia), which can aggra-
vate ICP related to vasogenic edema and break-
through of autoregulation [33]. Accordingly these
monitors can be used to optimize therapy: mannitol
and vasopressor infusion reduce ICP and improve
cerebral oxygenation when SjvO2 or PbtO2 values
fall below critical levels [30±32], whereas hyper-
ventilation reduces ICP and tissue oxygenation
when it is supranormal due to relative hyperperfu-
sion [31] (Table 4). The depth and duration of
ischemia detected by either device is highly corre-
lated with poor clinical outcome in patients with
severe TBI [30,31]. Since neither monitor alone can
detect all episodes of ischemia [34], selection of
which type to use depends on the speci®c clinical
circumstances at hand. PbtO2 monitoring is gener-
ally easier to use, and is most desirable when
detection of ischemia in a speci®c brain region is the
Table 4. Normal and Critical Values for SjvO2 and PbtO2
Monitors
SjvO2 PbtO2
(%) (mm Hg)
Normal valuea 62 37
Critical upper limit 80 NA
(indicates hyperemia)
Critical lower limit 50 8
(indicates ischemia)
Average value in severe TBIa 73 ‹ 10 32 ‹ 19
NA ˆ data not available.
Interpretation of critical values assumes normal FiO2 (£40%) and
hematocrit.
a
Data from reference 34.
predominant concern, whereas SjvO2 is less in¯u-
enced by high FiO2 levels and hence may be a more
reliable measure of relative hyperperfusion [34].
CEREBRAL MICRODIALYSIS. This adjunctive monitoring
technique is labor intensive and has yet to gain
widespread acceptance. A probe is placed through
the skull and levels of different substances may be
measured using high-performance liquid chromato-
graphy (HPLC). Lactate, glutamate, and more
recently extracellular potassium have been meas-
ured using this microdialysis. These levels correlate
with cerebral ischemia and poor outcome [35].
MULTIMODAL MONITORING. Finally, a combination of
multiple monitoring techniques may soon be
possible. Various prototypes are in development
that can allow simultaneous monitoring of ICP,
PbtO2, PbtCO2, pH, brain temperature, laser Dop-
pler ¯ow, and even microdialysis.
Management of Increased ICP
General Care Issues. Proper management of all
critically ill brain-injured patients begins with
general care issues designed to optimize oxygen-
ation and cerebral blood ¯ow and to minimize
factors that can aggravate neuronal injury or trigger
ICP elevations. The following guidelines should be
followed in all patients at risk for increased ICP:
PATIENT POSITIONING. As long as the patient is not
hypotensive (mean blood pressure < 60 mmHg),
the head of the bed should be elevated to 30
degrees and a straight head position should be
maintained. Head elevation to 30 degrees reduces
ICP by reducing jugular and cerebral venous
pressure and enhancing venous out¯ow, without
signi®cantly lowering CPP, CBF, or cardiac output
[36,37]. Some have recommended a head ¯at
position to prevent any reductions of CPP that

may occur with head elevation, which may be
reasonable in selected cases [38]. However, head
elevation in excess of 45 degrees should generally
be avoided because paradoxical increases in ICP
can occur in response to excessive CPP reduction
[39]. Sharp head angulation should also be avoided,
as it may cause jugular venous compression,
increased venous backpressure, and increased ICP.
FLUID MANAGEMENT. Only isotonic ¯uids, such as
0.9% (normal) saline or lactated Ringer's solution,
should be used in patients at risk for elevated ICP.
Cerebral edema is generally understood to result
from the creation of ``idiogenic osmoles'' which
draw water down an osmolar gradient into injured
brain tissue. Hypotonic ¯uids such as 5% dextrose
or 0.45% (half-normal) saline should be strictly
avoided because the free water contained in these
¯uids can aggravate cerebral edema and increase
ICP [40]. Systemic hypo-osmolality (<280 mOsm/L)
should be aggressively treated with mannitol or 3%
hypertonic saline. The traditional practice of
restricting total ¯uid intake (dehydration therapy),
with the goal of reducing the extracellular ¯uid
volume, has not been shown to signi®cantly impact
brain water content or ICP. In fact, hypovolemia
may lead to inadequate CPP and a consequent
increase in ICP [41].
Patients with elevated ICP should have a central
venous line placed to monitor central venous
pressure; this is particularly critical for patients
treated with mannitol or hypertonic saline. As a
general rule, a central venous pressure greater than
5 mmHg and equal to slightly positive daily net ¯uid
balance should be maintained by increasing the rate
of infusion of isotonic crystalloid, administering
0.9% saline or 5% albumin ¯uid boluses, or
transfusing blood to maintain hematocrit at greater
than 24% [42].
Use of 1.5±3% hypertonic saline as a maintenance
intravenous ¯uid for patients with cerebral edema is
gaining popularity in the neurocritical care com-
munity. Though support for this practice to date is
inconclusive, adequate clinical trials have yet to be
performed [43]. A small study of pediatric TBI
patients compared the use of 1.6% saline with
lactated Ringer's solution and found no differences
in ICP or CPP, though the hypertonic saline group
required fewer interventions to lower ICP [44]. Lack
of uniform concentrations and poor de®nition of
dose-response relationships has limited the wide-
spread use of hypertonic saline to date.
TEMPERATURE MANAGEMENT. Fevers should be trea-
ted aggressively. Temperature elevations increase
ICP by increasing cerebral metabolism and blood
Mayer and Chong: Management of ICP 61
¯ow, and have been shown to exacerbate hypoxic-
ischemic neuronal injury in experimental animals
[45]. As a general standard, acetaminophen and
cooling blankets should be given to all patients with
sustained fevers in excess of 38.3°C (101.0°F), but
evidence for their ef®cacy in neurologic patients is
scant [46,47]. Endovascular cooling with the use of
closed-circuit water-circulating intravenous cathe-
ters is a promising new approach that is currently
under development.
Recent studies suggest that indomethacin may be
the ideal antipyretic to use in patients with
increased ICP. Indomethacin has been shown to
decrease CBF and ICP in animal models and
patients with TBI [48]. The mechanism of action is
not known, but may involve vasoconstriction of
cerebral vessels and inhibition of prostaglandin
synthesis [48].
SEIZURE PROPHYLAXIS. Seizures can lead to profound
elevations of CBF, CBV, and ICP, even in patients
who are sedated or paralyzed [49]. This is secondary
to the increased cerebral metabolic demand that
occurs with seizures. Intravenous fosphenytoin
(15±20 mg/kg loading dose, 3±5 mg/kg/day) is
the preferred agent for seizure prophylaxis while in
the ICU.
STEROIDS. Dexamethasone and other steroids
should not be used as a standard treatment for ICP
because they are ineffective against cytotoxic
edema [50]. There is generally no role for steroids
in the treatment of mass effect related to cerebral
infarction [51], intracerebral hemorrhage [52], or TBI
[53]. By contrast, vasogenic edema related to
neoplasm or abscess is steroid responsive, and
dexamethasone 4±20 mg every 6 hours can lead to
dramatic reductions in lesion volume [54].
Emergent Treatment of Increased ICP in an
Unmonitored Patient. Emergency measures for
ICP control (Table 5) are appropriate for comatose
patients who present acutely with clinical signs of
increased ICP or herniation. These measures are
Table 5. Emergency Measures for ICP Reduction
1. Elevate head of bed 15±30 degrees.
2. Normal saline (0.9%) at 80±100 cc/hr
(avoid hypotonic ¯uids).
3. Intubate and hyperventilate (target pCO2 26±30
mmHg).
4. Mannitol 20% 1±1.5 g/kg via rapid intravenous
infusion.
5. Foley catheter.
6. CT scan and urgent neurosurgical consultation.
62 Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002
designed to lower ICP as quickly and effectively as
possible in order to ``buy time'' before a de®nitive
neurosurgical procedure (craniotomy, ventriculos-
tomy, or placement of an ICP monitor) can be
performed. Aggressive hyperventilation and mann-
itol therapy are the cornerstones of this type of
intervention.
Stepwise Treatment Protocol for Elevated ICP
in a Monitored Patient. The primary goal of ICP
management in a monitored patient is to maintain
ICP below 20 mmHg and CPP above 70 mmHg.
Contemporary ICP management has changed in
recent years in two important aspects: CPP man-
agement (in addition to ICP control) has become
increasingly emphasized, and the potential for
overzealous hyperventilation to cause excessive
vasoconstriction and aggravate ischemia has
become increasingly recognized. The stepwise
protocol presented here for managing ICP in a
monitored patient (Table 6) re¯ects these consid-
erations. More than one step may be instituted
simultaneously, but only after all of the preceding
steps have been addressed. Likewise, ICP therapy
should be withdrawn in a similar stepwise fashion
(Fig 7). This algorithm should be initiated any time
ICP remains greater than 20 mmHg for more than 10
minutes.
Though we favor the protocol presented below
because of its simplicity, other treatment algorithms
may be equally effective. Use of a standardized,
evidence-based protocol for ICP management al-
lows for ef®cient and well-coordinated treatment of
patients within an institution, and can improve
patient outcomes. In a retrospective study of
patients with TBI, patients treated with a standard-
ized algorithm of instituting and weaning ICP
therapies required fewer interventions and had a
Table 6. Stepwise Treatment Protocol for Elevated ICPa
in a Monitored Patient
1. Surgical decompression. Consider repeat CT
scanning and de®nitive surgical intervention or
ventricular drainage.
2. Sedation. Intravenous sedation to attain a
motionless, quiet state.
3. CPP optimization. Pressor infusion if CPP is less than
70 mmHg, or reduction of blood pressure if CPP is
greater than 110 mmHg.
4. Osmotherapy. Mannitol 0.25±1.0 g/kg intravenously
(repeat every 1±6 hours as needed).
5. Hyperventilation. Target pCO2 levels of 26±30 mmHg.
6. High-dose pentobarbital therapy. Load with 5±20
mg/kg, infuse 1±4 mg/kg/hr.
7. Hypothermia. Cool core body temperature to 32±33°C.
a
More than 20 mmHg for more than 10 minutes.
Refer to text for details.
Fig. 7. The Columbia stepwise protocol for ICP
management.
shorter duration of treatment than patients treated
ad hoc [55].
STEP 1: SURGICAL DECOMPRESSION OR CSF DRAIN-
AGE. The ®rst consideration in the face of an acute
increase in ICP should always be whether a
de®nitive intervention, such as craniotomy or
ventriculostomy, should be performed to remove
volume or decompress the skull. A repeat CT scan
should be considered to rule out reaccumulation of
an intracranial hemorrhage or worsening hydro-
cephalus. If a ventricular catheter is in place, the
system should be opened to drainage and 5±10 ml
of CSF removed.
The option of some de®nitive surgical interven-
tion should be continuously evaluated as additional
steps to control ICP are added. Controlled lumbar
CSF drainage (5±20 ml/hr) has been reported to
reduce ICP and increase CPP in patients refractory
to medical therapy and ventricular drainage alone
[56]. However, this intervention is feasible only if
the basal cisterns are open on CT, and even in these
cases, transtentorial herniation remains a risk.
Decompressive hemicraniectomy is becoming in-
creasingly used as an intervention of last resort for
patients who might otherwise require pentobarbital
or hypothermia. Wide cranial decompression can
de®nitively control ICP and reverse brain stem
herniation, and has been reported to be effective for
treating massive cerebral infarction [57], encephal-
itis [58], head trauma [59], and intracerebral hemor-
rhage [60] in nonrandomized studies. Complica-
tions of hemicraniectomy can include CSF leakage,
local wound infection or meningitis, intracranial
bleeding, and late hydrocephalus.
STEP 2: SEDATION. Sedation is often overlooked as a
key factor in ICP control. In patients with reduced
intracranial compliance, ®ghting against physical
restraints or ``bucking'' the ventilator can increase

ICP by elevating intrathoracic and jugular venous
pressure. Arterial hypertension associated with
agitation may further increase ICP if the patient is
at the upper range of the autoregulatory curve.
Before further measures are instituted, agitated
patients with increased ICP should be sedated to the
point where they are quiet and motionless (Ramsey
level 5 or 6). A combination of a sedative-hypnotic
and analgesic agent is usually most effective
(Table 7). The preferred regimen is the combina-
tion of an opioid, such as fentanyl (1±3 lg/kg/hr) or
sufentanil (0.1±0.6 lg/kg/hr), to provide analgesia
and propofol (0.3±3 mg/kg/hr) for sedation. It is
important to use drugs that are short acting, such
that the agent may be stopped for frequent neuro-
logic assessments throughout the day. One study
showed that daily, scheduled interruption of seda-
tion to examine patients not only reduced the
length of ventilator dependence and ICU length of
stay, but also decreased the need for other tests such
as brain imaging and lumbar punctures to evaluate
alterations in mental status [61].
Neurocritical care patients, even when comatose,
can sense pain and require analgesia in addition to
sedation. Therefore careful use of a low-dose con-
tinuous infusion opioid in addition to propofol or
midazolam is recommended. Bolus injections of
opioids, however, should be used with caution in
patients with elevated ICP. Bolus infusions of the
sufentanil, fentanyl, and alfentanil can transiently
lower MAP and increase ICP due to autoregulatory
vasodilation of cerebral vessels [62]. This effect is
seen primarily in patients with intact autoregulation,
but can also occur in patients with abnormal auto-
regulation [63]. Vasopressors may be used to avoid
hypotension and possible re¯ex ICP elevation.
Propofol may be the ideal sedative to use in
patients with elevated ICP; besides the fact that it is
Table 7. Selected Short-Acting Intravenous Sedatives for ICP Management
Agent Pharmacology
Sedative-analgesic agents
Morphine sulfate Opioid (sedative-hypnotic with analgesic
properties)
Fentanyl Opioid (short acting, 100 times more potent than
morphine)
Sufentanil Opioid (ultrashort acting)
Sedative-hypnotic agents
Propofol Alkylphenol (ultrashort acting)
Midazolam Benzodiazepine (short acting)
Dosages are approximate and should be titrated to the patient's level of agitation and ICP. A combination of a sedative-analgesic and
sedative-hypnotic agent may be more effective than the use of a single agent.
Mayer and Chong: Management of ICP 63
ultrashort acting, it has favorable effects on ICP and
seizure activity, and may have neuroprotective
properties. In a study comparing propofol to
morphine in patients with severe TBI, patients
treated with propofol had lower ICP values and
more favorable long-term neurologic outcomes
[64]. By contrast, paralysis with neuromuscular
blocking agents such as vecuronium, pancuronium,
or cis-atracurium is rarely necessary, and places
patients at risk for prolonged paralysis due to
critical illness myopathy.
STEP 3: CPP OPTIMIZATION. If CPP is less than 70
mmHg and ICP is greater than 20 mmHg, elevation
of mean arterial blood pressure and CPP with a
vasopressor such as dopamine or phenylephrine
(Table 8) can lead to a re¯ex reduction of ICP, by
eliminating cerebral vasodilation that occurs in
response to inadequate perfusion (Fig 8). CPP
optimization to levels well above 70 mmHg may
be desirable in chronically hypertensive patients
(whose autoregulatory curve is shifted to the right),
or in patients with low PbtO2 or SjvO2 levels or
Lundberg A and B waves (since these ®ndings
generally re¯ect insuf®cient CPP). The widely
accepted ``one size ®ts all'' approach to CPP
management (>70 mmHg) is in all likelihood an
oversimpli®cation, and efforts should be made to
optimize CPP whenever patients fail standard
therapy. One study that attempted to de®ne optimal
CPP levels for severe TBI patients by analyzing
receiver-operating curves found that a CPP of 55
mmHg was the critical threshold for poor outcome
in adults [65]. It seems prudent to maintain CPP well
above this level in clinical practice, however, to
provide an extra margin of safety. In an uncon-
trolled study of TBI patients with CPP main-
tained above 70 mmHg with phenylephrine and
Dosage Range
2±5 mg IVP every 1±4 hours
0.5±3.0 lg/kg/hr
0.1±0.6 lg/kg/hr
0.6±6 mg/kg/hr
0.05±0.1 mg/kg/hr
64 Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002
Table 8. Selected Short-Acting Vasoactive Drugs for ICP
Management
Agent Pharmacology Dosage Range
Blood pressure reduction
Labetolol a1 and b1 blocker 2±3 mg/min
Nicardipine Calcium channel 5±15 mg/hr
blocker
Blood pressure elevation
Dopamine a1 and b1 agonist 5±30 lg/kg/min
(at high doses)
Norepinephrine a1 and b1 agonist 0.03±0.6 lg/kg/min
Phenylephrine a1 agonist 2±10 lg/kg/min
norepinephrine, better outcomes were obtained
than in previously reported patients with strictly
ICP-based management [41].
If MAP and ICP are elevated in a sedated patient,
treatment of arterial hypertension can sometimes
lead to a parallel reduction of ICP. If CPP is greater
than 110 mmHg and ICP is greater than 20 mmHg,
hypertension should be carefully treated with a
short-acting agent (Table 7). However, extreme
caution should be used to avoid reduction of CPP
below 70 mmHg, which can trigger re¯ex cerebral
vasodilation and ICP elevation. Nitroprusside can
be particularly troublesome in this regard, and
Fig 8. Demonstration of CPP augmentation resulting in
re¯ex ICP reduction. At approximately 9:30, the patient
developed a plateau wave with ICP elevation from 20 to
40 mmHg, and CPP reduction from 100 to 70 mmHg. A
dopamine infusion was started at 9:45, with nearly im-
mediate normalization of ICP to 20 mmHg when CPP was
restored to 100 mmHg.
concerns also exist regarding the potential for
nitroprusside to directly dilate the cerebral vascu-
lature and increase ICP. For these reasons, nitro-
prusside is best avoided in patients with elevated
ICP, and should only be used if an ICP monitor is in
place. Experimental and clinical studies have dem-
onstrated that treatment of hypertension in patients
with intracerebral hemorrhage does not exacerbate
perilesional ischemia [66] or cause re¯ex vasodila-
tion and increased ICP [67], as long as CPP remains
within the normal range.
STEP 4: OSMOTHERAPY. If CPP is optimized, the patient
is sedated, and ICP remains elevated, mannitol or
hypertonic saline should be given. Mannitol, an
osmotic diuretic, lowers ICP via its cerebral dehy-
drating effects. The effects of mannitol are biphasic.
Rapid infusion immediately creates an osmotic
gradient across the blood-brain barrier, which leads
to movement of water from brain parenchyma
to the intravascular compartment. The result is
decreased brain tissue volume, and hence reduced
ICP [68]. The secondary effect of mannitol results
from its action as an osmotic diuretic. As mannitol is
cleared by the kidneys, it leads to free water
clearance and increased serum osmolality. This
leads to a more prolonged intracellular dehydrating
effect as water ¯ows down the osmotic gradient
from the intracellular to the extracellular space.
The initial dose of mannitol 20% solution is 1±1.5
g/kg, followed every 1±6 hours with doses of 0.25±1
g/kg as needed. Repeated mannitol doses should
be given on the basis of ICP measurements rather
than as scheduled doses, unless the goal is to
establish and maintain a hyperosmolar state (300±
320 mOsm/L). The effect of a mannitol bolus on ICP
begins within 10±20 minutes, reaches its peak
between 20 and 60 minutes, and lasts for 4±6 hours,
but this may vary widely between patients. Adverse
effects of mannitol therapy include exacerbation of
congestive heart failure (due to the initial intravas-
cular volume expansion); volume contraction,
hypokalemia, and profound hyperosmolality (after
prolonged use); acute tubular necrosis (due to
excessive hyperosmolality); and ``rebound'' increa-
ses in ICP [69]. Patients treated repeatedly with
mannitol require frequent measurements of serum
electrolytes and osmolality, careful recording of
¯uid input and output, and central venous pressure
monitoring. Urinary volume losses should be
replaced with normal (0.9%) saline to avoid volume
depletion.
Bolus infusion of 3%, 7.5%, 10%, or 24.3%
hypertonic saline is gaining popularity as an
alternative to bolus infusions of mannitol for ICP
crises [43]. In an uncontrolled study of 3% saline/

acetate infusion (75±150 ml/hr) and intermittent
boluses (250 ml) titrated to maintain serum sodium
levels between 145 and 155 mEq/L, hypertonic
saline lowered ICP related to TBI and brain tumors,
but not in patients with intracerebral hemorrhage or
cerebral infarction [70]. Infusion of 2±5 ml/kg of
7.5% saline or 0.5±1.0 ml/kg of 23.4% saline over 30
minutes has also been shown to lower elevated ICP
and augment CPP, with an effect that lasts several
hours [71,72]. A large head-to-head trial of mannitol
versus hypertonic saline is warranted.
STEP 5: HYPERVENTILATION. As a general rule, the goal
of hyperventilation for ICP control should be to
lower pCO2 to 30 mmHg, or to 25±30 mmHg in
extreme cases. The respiratory alkalosis caused by
hyperventilation lowers ICP by causing cerebral
vasoconstriction and reduced CBV [1]. The peak
effect of hyperventilation on ICP is generally
reached within 30 minutes. Over the next 1±3 hours
the effect gradually diminishes, as compensatory
acid-base buffering mechanisms correct the alkalo-
sis within the central nervous system [1]. In patients
with elevated ICP due to hyperemia and increased
CBV, however, the effect may be prolonged, and
hyperventilation may be the treatment of choice
[73]. Hyperventilation should be tapered slowly
over 4±6 hours because abrupt cessation may lead
to vasodilation and rebound increases in ICP.
Though it is postulated that prolonged severe
hyperventilation (pCO2 < 25 mmHg) can actually
exacerbate cerebral ischemia by causing profound
vasoconstriction, the deleterious effects of exces-
sive hyperventilation may also include more labile
ICP. In a study of patients with TBI, prophylactic
hyperventilation to a pCO2 of 25 mmHg resulted in
poorer outcome compared with normally ventila-
ted patients [74]. The authors hypothesized that
blood vessels may become hypersensitive to chan-
ges in pCO2 after prolonged hyperventilation
because CSF buffering capacity is lost. In cases of
pediatric head trauma, profound hypocarbia was
associated with decreased cerebral oxygen con-
sumption and ischemia [75]. SjvO2 monitoring is a
useful modality for ensuring that prolonged
aggressive hyperventilation, if necessary, is not
critically reducing oxygen delivery to the brain.
STEP 6: PENTOBARBITAL THERAPY. High-dose barbitur-
ate therapy, given in doses equivalent general
anesthesia, can effectively lower ICP in most
patients refractory to the steps outlined above
[76,77]. The effect of pentobarbital is multifactorial,
but most likely stems from a profound reduction of
cerebral metabolism, which is coupled to reduc-
tions of CBF and CBV [78]. Pentobarbital often
Mayer and Chong: Management of ICP 65
causes hypotension, and usually requires the use of
vasopressors to maintain CPP above 70 mmHg.
Pentobarbital typically requires a loading dose of
10±20 mg/kg, given in repeated 5 mg/kg boluses,
until a state of ¯accid coma with preserved pupillary
reactivity is attained. Maintenance infusion is usu-
ally about 1±4 mg/kg/hr. Continuous EEG monit-
oring is helpful to avoid oversedation, since gener-
ally no further ICP reduction occurs once a burst-
suppression pattern is attained. If ICP is normalized
with pentobarbital, it is generally maintained for
24±48 hours. It can then be abruptly discontinued
because its highly lipophilic nature and long half-
life (90 hours) result in a gradual reduction of blood
levels over several days.
STEP 7: HYPOTHERMIA. Systemic hypothermia to lev-
els of 32±33°C can lower ICP in some patients
refractory to pentobarbital [79±81]. This technique
requires placement of cooling blankets under and
over the patient, iced gastric lavage, and pharma-
cologic paralysis with vecuronium or a similar
neuromuscular blocking agent to prevent shiver-
ing. Prolonged hypothermia can be dangerous
because of increased risk of infectious complica-
tions, coagulopathy, and electrolyte derangements,
among other hazards. Rewarming should always be
done slowly, over at least a day, and passively,
without active heating, to avoid rebound cerebral
edema or a systemic in¯ammatory response syn-
drome, which can be fatal.
In a small randomized controlled trial of severe
TBI patients refractory to pentobarbital, mild-to-
moderate hypothermia (34°C) signi®cantly reduced
ICP, improved CPP, reduced CBF and cerebral
metabolic rate (CMRO2), and reduced arteriojugu-
lar venous oxygen differences [79]. Survival was
50% in hypothermia patients compared to 18% in
the control group (p < 0.05). Later studies by the
same group reported that hypothermia was most
effective for pentobarbital-refractory ICP elevations
between 20 and 40 mmHg [80], and that severe TBI
patients with low ICP do not bene®t from hypo-
thermia [81]. These reports, and the negative results
of a recent large National Institutes of Health (NIH)-
funded trial studying the effects of hypothermia as
®rst-line therapy for severe TBI [82], suggest that
hypothermia is not effective as a primary form of
neuroprotection for severe TBI related to diffuse
axonal injury.
References
1. Ropper AH, Rockoff MA. Physiology and clinical aspects of
raised intracranial pressure. In: Ropper AH, ed. Neurological
66 Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002
and neurosurgical intensive care, 3rd ed. Philadelphia: Lip-
pincott-Raven, 1993:11±28
2. Jennet WB, Harper AM, Meller JD, et al. Relationship be-
tween cerebral blood ¯ow and cerebral perfusion pressure.
Br J Surg 1970;57:390±397
3. Shinnoj E, Strangaard S. Pathogenesis of hypertensive en-
cephalopathy. Lancet 1973;1:461±462
4. Nates JL, Niggemeyer LE, Anderson MB, et al. Cerebral per-
fusion pressure monitoring alert! [letter]. Crit Care Med
1997;25:895±896
5. Lundberg N. Continuous recording and control of ventricular
¯uid pressure in neurosurgical practice. Acta Psychiatr Scand
Suppl 1960;149:1±19
6. Rosner MJ, Becker DP. Origin and evolution of plateau
waves: experimental observations and a theoretical model.
J Neurosurg 1984;60:312±324
7. Czosnyka M, Smielewski P, Piechnik S, et al. Hemodynamic
characterization of intracranial pressure plateau waves in
head-injured patients. J Neurosurg 1999;91:11±19.
8. Ropper AH, King RB. Intracranial pressure monitoring in
comatose patients with cerebral hemorrhage. Arch Neurol
1984;41:725
9. Frank J. Large hemispheric infarction, deterioration and in-
tracranial pressure. Neurology 1995;45:1286±1290
10. Wol¯a CE, Luerssen TG, Bowman RM, et al. Brain tissue
pressure gradients by expanding frontal epidural mass
lesion. J Neurosurg 1996;84:642±647
11. Sahuquillo J, Poca M-A, Arribas M, et al. Intrahemispheric
supratentorial intracranial pressure gradients in head-injured
patients: Are they clinically important? J Neurosurg
1999;90:16±26
12. Narayan RK, Kishore PR, Becker DP, et al. Intracranial
pressure: To monitor or not to monitor? A review of our
experience with severe head injury. J Neurosurg
1982;56:650±659
13. Papo I, Janny P, Caruselli G, et al. Intracranial pressure time
course in primary intracerebral hemorrhage. Neurosurgery
1979;4:504±511
14. Mayhall CG, Archer NH, Lamb VA, et al. Ventriculostomy
related infections: a prospective epidemiologic study. N Engl
J Med 1984;310:553±559
15. Poon WS, Ng S, Wai S. CSF antibiotic prophylaxis for neu-
rosurgical patients with ventriculostomy: a randomised
study. Acta Neurochir Suppl 1998;71:146±148
16. Ostrup RC, Luerssen TG, Marshall LF, et al. Continuous
monitoring of intracranial pressure with a miniaturized
®beroptic device. J Neurosurg 1987;67:206±209
17. Gopinath SP, Robertson CS, Contant CF, et al. Clinical eval-
uation of a miniature strain-gauge transducer for monitoring
intracranial pressure. Neurosurgery 1995;36:1137±1140
18. Levin A. The use of a ®beroptic intracranial pressure monitor
in clinical practice. Neurosurgery 1977;1:266±271
19. Sundbarg G, Nordstrom CH, Messeter K, et al. A com-
parison of intraparenchymatous and intraventricular pres-
sure recording in clinical practice. J Neurosurg
1987;67:841±845
20. Yau Y-H, Piper IR, Clutton RE, et al. Experimental evaluation
of the Spielberger intracranial pressure and intracranial
compliance monitor. Technical note. J Neurosurg
2000;93:1072±1077
21. Vries JK, Becker DP, Yang HF. A subarachnoid screw for
monitoring intracranial pressure. J Neurosurg 1973;39:416±
419
22. Miller JD, Bobo H, Kapp JP. Inaccurate pressure readings
from subarachnoid bolts. Neurosurgery 1986;19:253±255
23. Ream AK, Silverberg GD, Corbin SD, et al. Epidural meas-
urement of intracranial pressure. Neurosurgery 1979;5:36±43
24. Hassler W, Steinmetz H, Gawlowski J. Transcranial Doppler
ultrasonography in raised intracranial pressure and in intra-
cranial circulatory arrest. J Neurosurg 1988;68:745±751
25. Mayer SA, Thomas CE, Diamond BE. Asymmetry of intra-
cranial hemodynamics as an indicator of mass effect in acute
intracerebral hemorrhage: a transcranial Doppler study.
Stroke 1996;27:1788±1792
26. Schmidt B, KlingelhoÈfer J, Schwarze JJ, et al. Noninvasive
prediction of intracranial pressure curves using transcranial
Doppler ultrasonography and blood pressure curves. Stroke
1997;28:2465±2472
27. Pranevicius O, Bertasius K, Pranevicius K, et al. Non-inva-
sive dynamic assessment of the elasticity of intracranial
structures. Acta Neurol Scand 1992;86:512±516
28. Czosnyka M, Smielewski P, Piechnik S, et al. Critical closing
pressure in cerebrovascular circulation. J Neurol Neurosurg
Psychiatry 1999;66:606±611
29. Robertson C, Gopinath S, Goodman JC et al. SjvO2 monit-
oring in head-injured patients. J Neurotrauma 1995;12:891±
896
30. van den Brink WA, van Santbrink H, Steyerberg EW, et al.
Brain oxygen tension in severe head injury. Neurosurgery
2000;46:868±878
31. Cruz J. The ®rst decade of continuous monitoring of jugular
bulb oxyhemoglobin saturation: management strategies and
clinical outcome. Crit Care Med 1998;26:344±351
32. Cruz J. Continuous monitoring of cerebral oxygenation in
acute brain injury: injection of mannitol during hyperventi-
lation. J Neurosurg 1990;73:725±730
33. Cormio M, Valadka AB, Robertson CS. Elevated jugular ve-
nous oxygen saturation after severe head injury. J Neurosurg
1999;90:9±15
34. Gopinath SP, Valadka AB, Uzura M, et al. Comparison of
jugular venous oxygen saturation and brain tissue PO2 as
monitors of cerebral ischemia after head injury. Crit Care Med
1999;27:2337±2345
35. Hutchinson P, Hutchinson D, Barr RH et al. A new cranial
access device for cerebral monitoring. Br J Neurosurg
2000;14:46±48
36. Durward QJ, Amacher AL, DelMaestro RF, et al. Cerebral
and cardiovascular responses to head elevation in patients
with intracranial hypertension. J Neurosurg 1983;59:938±
944
37. Feldman Z, Kanter MJ, Robertson CS, et al. Effect of head
elevation on intracranial pressure, cerebral perfusion pres-
sure, and cerebral blood ¯ow in head-injured patients. J
Neurosurg 1992;76:207±211
38. Rosner MJ, Coley IB. Cerebral perfusion pressure, intra-
cranial pressure, and head elevation. J Neurosurg
1986;65:636±641
39. Moraine J, Berre J, Melot C. Is cerebral perfusion pressure a
major determinant of cerebral blood ¯ow during head ele-
vation in comatose patients with severe intracranial lesions? J
Neurosurg 2000;92:606±614
40. Mayer SA. Fluid management in subarachnoid hemorrhage.
Neurologist 1995;1:71±85.
41. Rosner MJ, Rosner SD, Johnson AH. Cerebral perfusion
pressure: management protocol and clinical results. J Neu-
rosurg 1995;83:949±962
42. Herbert PC, Wells G, Blajchman MA, et al. A multicenter,
randomized, controlled clinical trial of transfusion require-
ments in critical care. N Engl J Med 1999;340:409±417
43. Qureshi AI, Suarez JI. Use of hypertonic saline solutions in
treatment of cerebral edema and intracranial hypertension.
Crit Care Med 2000;28:3301±3313
44. Simma B, Burger R, Falk M, et al. A prospective, randomized
and controlled study of ¯uid management in children with

severe head injury; lactated Ringer's solution versus hyper-
tonic saline. Crit Care Med 1998;26:1265±1270
45. Busto R, Dietrich WD, Globus MY, et al. Small differences in
intraischemic brain temperature critically determine the ex-
tent of ischemic injury. J Cereb Blood Flow Metab
1987;7:129±138
46. O'Donnell J, Axelrod P, Fisher C, et al. Use and effectiveness
of hypothermia blankets for febrile patients in the intensive
care unit. Clin Infect Dis 1996;24:1208±1213
47. Mayer SA, Commichau C, Scarmeas N, et al. Clinical trial of
an air-circulating cooling blanket for fever control in critic-
ally-ill neurologic patients. Neurology 2001;56:292±298
48. Slavik R, Rhoney D. Indomethacin: a review of its cerebral
blood ¯ow effects and potential use for controlling intra-
cranial pressure in traumatic brain injury patients. Neurol Res
1999;21:491±499
49. Lassen NA. Control of the cerebral circulation in health and
disease. Circ Res 1974;34:749±760
50. Fishman RA. Brain edema. N Engl J Med 1975;293:706±711
51. Anderson DC, Cranford RE. Corticosteroids in ischemic
stroke. Stroke 1979;10:68±71
52. Pourgvarin H, Bhoopat TW, Viriyavejakul A, et al. Effects of
dexamethasone in primary supratentorial intracerebral
hemorrhage. N Engl J Med 1987;316:1229±1233
53. Cooper PR, Moody S, Clark WK, et al. Dexamethasone and
severe head injury: a prospective double-blind trial. J Neu-
rosurg 1979;51:307±331
54. Galich JM, French LA. Use of dexamethasone in the treatment
of cerebral edema resulting from brain tumors and brain
surgery. Am Pract 1961;12:169±174
55. McKinley B, Parmley C, Tonneson A. Standardized man-
agement of intracranial pressure: a preliminary clinical trial. J
Trauma 1999;46:271±279
56. MuÈnch EC, Bauhuf C, Horn P, et al. Therapy of malignant
intracranial hypertension by controlled lumbar cerebro-
spinal ¯uid drainage. Crit Care Med 2001;29:976±981
57. Schwab S, Steiner T, Aschoff A, et al. Early hemicraniectomy
in patients with complete middle cerebral artery infarction.
Stroke 1998;29:1888±1893
58. Schwab S, Junger E, Spranger M, et al. Craniectomy: an
aggressive treatment approach in severe encephalitis.
Neurology 1997;48:412±417
59. Guerra WK, Gaab MR, Dietz H, et al. Surgical decompres-
sion for traumatic brain swelling: indications and results. J
Neurosurg 1999;90:187±196
60. Mayer SA, Connolly ES, Bates J, et al. Decompressive hemi-
craniectomy for massive cerebral hemorrhage [abstract].
Stroke 2001;32:357
61. Kress JP, Pohlman AS, O'Connor MF, et al. Daily inter-
ruption of sedative infusions in critically ill patients
undergoing mechanical ventilation. N Engl J Med 2000;
342:1471±1477
62. Albanese J, Viviand X, Potie F, et al. Sufentanil, fentanyl, and
alfentanil in head trauma patients: a study on cerebral
hemodynamics. Crit Care Med 1999;27:407±411
63. de Nadal M, Ausina A, et al. Effects on intracranial pressure of
fentanyl in severe head injured patients. Neurochirurgia
1998;71(suppl):10±12
64. Kelly D, Goodale D, Williams J, et al. Propofol in the treat-
ment of moderate and severe head injury: a randomized,
prospective double-blinded pilot trial. J Neurosurg
1999;90:1042±1051
Mayer and Chong: Management of ICP 67
65. Chambers I, Treadwell R, Mendelow AD. Determination of
threshold levels of cerebral perfusion pressure and intra-
cranial pressure in severe head injury by using receiver-
operating characteristic curves: an observational study in
291 patients. J Neurosurg 2001;94:412±416
66. Powers WJ, Zazulia AR, Videen TO, et al. Autoregulation of
cerebral blood ¯ow surrounding acute (6±22 hours) intra-
cerebral hemorrhage. Neurology 2001;57:18±24
67. Qureshi AI, Wilson DA, Hanley DF, et al. Pharmacologic
reduction of mean arterial pressure does not adversely affect
regional cerebral blood ¯ow and intracranial pressure in
experimental intracerebral hemorrhage. Crit Care Med
1999;27:965±971
68. Messeter K, Nordstrom CH, Sundbarg G, et al. Cerebral
hemodynamics in patients with acute severe head injury.
J Neurosurg 1986;64:231±237
69. Kaufmann AM, Cardoso ER. Aggravation of vasogenic
cerebral edema by multiple dose mannitol. J Neurosurg
1993;44:584±589
70. Qureshi A, Suarez J, Bhardwaj A, et al. Use of hypertonic
(3%) saline/acetate infusion in the treatment of cerebral
edema: effect on intracranial pressure and lateral displace-
ment of the brain. Crit Care Med 1998;26:440±446
71. Suarez J, Qureshi A, Bhardwaj A, et al. Treatment of re-
fractory intracranial hypertension with 23.4% saline. Crit
Care Med 1998;26:1118±1122
72. Hartl R, Ghajar J, Hochleuthner, et al. Hypertonic/hyper-
oncotic saline reliably reduces ICP in severely head-injured
patients with intracranial hypertension. Acta Neurochir
Suppl 1997;70:126±129
73. Obrist W, Lang®tt T, Jaggi JL, et al. Cerebral blood ¯ow and
metabolism in comatose patients with acute head injury. J
Neurosurg 1984;61:241±253
74. Muizelaar JP, Marmarou A, Ward JD, et al. Adverse effects of
prolonged hyperventilation in patients with severe head
injury: a randomized clinical trial. J Neurosurg 1991;75:731±
739
75. Skippen P, Seear M, Poskitt K, et al. Effect of hyperventila-
tion on regional cerebral blood ¯ow in head-injured chil-
dren. Crit Care Med 1997;25:1402±1409
76. Rea GL, Rockswold GL. Barbiturate therapy in uncon-
trolled intracranial hypertension. Neurosurgery 1983;12:
401±404
77. Eisenberg HM, Frankowski RF, Contant CF, et al. High-dose
barbiturate control of elevated intracranial pressure in
patients with severe head injury. J Neurosurg 1988;69:15±23
78. Cormio M, Gopinath S, Valadka A, et al. Cerebral hemo-
dynamic effects of pentobarbital coma in head-injured
patients. J Neurotrauma 1999;16:927±936
79. Shiozaki T, Sugimoto H, Taneda M, et al. Effect of mild
hypothermia on uncontrollable intracranial hypertension
after severe head injury. J Neurosurg 1993;79:363±368
80. Shiozaki T, Sugimoto H, Taneda M, et al. Selection of
severely head injured patients for mild hypothermia therapy.
J Neurosurg 1998;89:206±211
81. Shiozaki T, Kato A, Taneda M, et al. Little bene®t from
mild hypothermia therapy for severely head injured
patients with low intracranial pressure. J Neurosurg 1999;
91:185±191
82. Clifton GL, Miller ER, Choi SC, et al. Lack of effect of induc-
tion of hypothermia after acute brain injury. N Engl J Med
2001;344:556±563